1.1 Field of the Invention
The present invention relates generally to the fields of insect control. Certain embodiments concern methods and compositions comprising nucleic acid segments which encode Bacillus thuringiensis-derived .delta.-endotoxins. Disclosed are methods of altering Cry1 crystal proteins by mutagenesis of the loop regions between the .alpha.-helices of the protein's domain 1 or of the loop region between .alpha.-helix 7 of domain 1 and .beta.-strand 1 of domain 2 to give rise to modified Cry1 proteins (Cry1*) which have improved activity against Lepidopteran insects. Various methods for making and using these recombinantly-engineered proteins and nucleic acid segments, including development of transgenic plant cells and recombinant host cells are also disclosed.
1.2 Description of the Related Art
The most widely used microbial pesticides are derived from the bacterium Bacillus thuringiensis. B. thuringiensis is a Gram-positive bacterium that produces crystal proteins which are specifically toxic to certain orders and species of insects. Many different strains of B. thuringiensis have been shown to produce insecticidal crystal proteins. Compositions including B. thuringiensis strains which produce insecticidal proteins have been commercially-available and used as environmentally-acceptable insecticides because they are quite toxic to the specific target insect, but are harmless to plants and other non-targeted organisms.
.delta.-endotoxins are used to control a wide range of leaf-eating caterpillars and beetles, as well as mosquitoes. B. thuringiensis produces a proteinaceous parasporal body or crystal which is toxic upon ingestion by a susceptible insect host. For example, B. thuringiensis subsp. kurstaki HD-1 produces a crystal inclusion comprising .delta.-endotoxins which are toxic to the larvae of a number of insects in the order Lepidoptera (Schnepf and Whiteley, 1981).
1.2.1 .delta.-ENDOTOXINS
.delta.-endotoxins are a large collection of insecticidal proteins produced by B. thuringiensis. Over the past decade research on the structure and function of B. thuringiensis toxins has covered all of the major toxin categories, and while these toxins differ in specific structure and function, general similarities in the structure and function are assumed. Based on the accumulated knowledge of B. thuringiensis toxins, a generalized mode of action for B. thuringiensis toxins has been created and includes: ingestion by the insect, solubilization in the insect midgut (a combination stomach and small intestine), resistance to digestive enzymes sometimes with partial digestion actually "activating" the toxin, binding to the midgut cells, formation of a pore in the insect cells and the disruption of cellular homeostasis (English and Slatin, 1992).
1.2.2 GENES ENCODING CRYSTAL PROTEINS
Many of the .delta.-endotoxins are related to various degrees by similarities in their amino acid sequences. Historically, the proteins and the genes which encode them were classified based largely upon their spectrum of insecticidal activity. The review by Hofte and Whiteley (1989) discusses the genes and proteins that were identified in B. thuringiensis prior to 1990, and sets forth the nomenclature and classification scheme which has traditionally been applied to B. thuringiensis genes and proteins. cryI genes encode lepidopteran-toxic CryI proteins. cryII genes encode CryII proteins that are toxic to both lepidopterans and dipterans. cryIII genes encode coleopteran-toxic CryIII proteins, while cryIV genes encode dipteran-toxic CryIV proteins.
Based on the degree of sequence similarity, the proteins were further classified into subfamilies; more highly related proteins within each family were assigned divisional letters such as CryIA, CryIB, CryIC, etc. Even more closely related proteins within each division were given names such as CryIC1, CryIC2, etc.
Recently a new nomenclature has been proposed which systematically classifies the Cry proteins based upon amino acid sequence homology rather than upon insect target specificities. This classification scheme is summarized in Table 1.
TABLE 1 REVISED B. THURINGIENSIS .delta.-ENDOTOXIN NOMENCLATURE.sup.A New Old GenBank Accession # Cry1Aa CryIA(a) M11250 Cry1Ab CryIA(b) M13898 Cry1Ac CryIA(c) M11068 Cry1Ad CryIA(d) M73250 Cry1Ae CryIA(e) M65252 Cry1Ba CryIB X06711 Cry1Bb ET5 L32020 Cry1Bc PEG5 Z46442 Cry1Bd CryE1 U70726 Cry1Ca CryIC X07518 Cry1Cb CryIC(b) M97880 Cry1Da CryID X54160 Cry1Db PrtB Z22511 Cry1Ea CryIE X53985 Cry1Eb CryIE(b) M73253 Cry1Fa CryIF M63897 Cry1Fb PrtD Z22512 Cry1Ga PrtA Z22510 Cry1Gb CryH2 U70725 Cry1Ha PrtC Z22513 Cry1Hb U35780 Cry1Ia CryV X62821 Cry1Ib CryV U07642 Cry1Ja ET4 L32019 Cry1Jb ET1 U31527 Cry1K U28801 Cry2Aa CryIIA M31738 Cry2Ab CryIIB M23724 Cry2Ac CryIIC X57252 Cry3A CryIIIA M22472 Cry3Ba CryIIIB X17123 Cry3Bb CryIIIB2 M89794 Cry3C CryIIID X59797 Cry4A CryIVA Y00423 Cry4B CryIVB X07423 Cry5Aa CryVA(a) L07025 Cry5Ab CryVA(b) L07026 Cry5B U19725 Cry6A CryVIA L07022 Cry6B CryVIB L07024 Cry7Aa CryIIIC M64478 Cry7Ab CryIIICb U04367 Cry8A CryIIIE U04364 Cry8B CryIIIG U04365 Cry8C CryIIIF U04366 Cry9A CryIG X58120 Cry9B CryIX X75019 Cry9C CryIH Z37527 Cry10A CryIVC M12662 Cry11A CryIVD M31737 Cry11B Jeg80 X86902 Cry12A CryVB L07027 Cry13A CryVC L07023 Cry14A CryVD U13955 Cry15A 34kDa M76442 Cry16A cbm71 X94146 Cry17A cbm71 X99478 Cry18A CryBP1 X99049 Cry19A Jeg65 Y08920 Cyt1Aa CytA X03182 Cyt1Ab CytM X98793 Cyt1B U37196 Cyt2A CytB Z14147 Cyt2B CytB U52043 .sup.A Adapted from: http://epunix.biols.susx.ac.uk/Home/Neil_Crickmore/Bt/index.html
1.2.3 CRYSTAL PROTEINS FIND UTILITY AS BIOINSECTICIDES
The utility of bacterial crystal proteins as insecticides was extended when the first isolation of a coleopteran-toxic B. thuringiensis strain was reported (Krieg et al., 1983; 1984). This strain (described in U.S. Pat. No. 4,766,203, specifically incorporated herein by reference), designated B. thuringiensis var. tenebrionis, is reported to be toxic to larvae of the coleopteran insects Agelastica alni (blue alder leaf beetle) and Leptinotarsa decemlineata (Colorado potato beetle).
U.S. Pat. No. 5,024,837 also describes hybrid B. thuringiensis var. kurstaki strains which showed activity against lepidopteran insects. U.S. Pat. No. 4,797,279 (corresponding to EP 0221024) discloses a hybrid B. thuringiensis containing a plasmid from B. thuringiensis var. kurstaki encoding a lepidopteran-toxic crystal protein-encoding gene and a plasmid from B. thuringiensis tenebrionis encoding a coleopteran-toxic crystal protein-encoding gene. The hybrid B. thuringiensis strain produces crystal proteins characteristic of those made by both B. thuringiensis kurstaki and B. thuringiensis tenebrionis. U.S. Pat. No. 4,910,016 (corresponding to EP 0303379) discloses a B. thuringiensis isolate identified as B. thuringiensis MT 104 which has insecticidal activity against coleopterans and lepidopterans.
1.2.4 CRY1 ENDOTOXINS
The characterization of the lepidopteran-toxic B. thuringiensis Cry1Aa crystal protein, and the cloning, DNA sequencing, and expression of the gene. which encodes it have been described (Schnepf and Whitely, 1981; Schnepf et al., 1985). In related publications, U.S. Pat. No. 4,448,885 and U.S. Pat. No. 4,467,036 (specifically incorporated herein by reference), the expression of the native B. thuringiensis Cry1Aa crystal protein in E. coli is disclosed.
Several cry1C genes have been described in the prior art. A cry1C gene truncated at the 3' end was isolated from B. thuringiensis subsp. aizawai 7.29 by Sanchis et al. (1988). The truncated protein exhibited toxicity towards Spodoptera species. The sequence of the truncated cry1C gene and its encoded protein was disclosed in PCT WO 88/09812 and in Sanchis et al., (1989). The sequence of a cry1C gene isolated from B. thuringiensis subsp. entomocidus 60.5 was described by Honee et al., (1988). This gene is recognized as the holotype cry1C gene by Hofte and Whiteley (1989). The sequence of a cry1C gene is also described in U.S. Pat. No. 5,126,133.
The cry1C gene from B. thuringiensis subsp. aizawai EG6346, contained on plasmids pEG315 and pEG916 described herein, encodes a Cry1C protein identical to that described in the aforementioned U.S. Pat. No. 5,126,133. The Cry1C protein described by Sanchis et al., (1989) and in PCT WO 88/09812 differs from the EG6346 Cry1C protein at several positions that can be described as substitutions within the EG6346 Cry1C protein: Cry1C N366I, W376C, P377Q, A378R, P379H, P380H, V386G, R775A.
Significantly, the amino acid positions 376-380 correspond to amino acid residues predicted to lie within the loop region between .beta. strand 6 and .beta. strand 7 of Cry1C, using the nomenclature adopted by Li et al. (1991) for identifying structures within Cry3A. Bioassay comparisons between the Cry1C protein of strain EG6346 and the Cry1C protein of strain aizawai 7.29 revealed no significant differences in insecticidal activity towards S. exigua, T. ni, or P. xylostella. These results suggested that the two Cry1C proteins exhibited the same insecticidal specificity in spite of their different amino acid sequences within the predicted loop region between .beta. strand 6 and .beta. strand 7.
Smith and Ellar (1994) reported the cloning of a cry1C gene from B. thuringiensis strain HD229 and demonstrated that amino acid substitutions within the putative loop region between .beta. strand 6 and .beta. strand 7 ("loop .beta. 6-7") altered the insecticidal specificity of Cry1C towards Spodoptera frugiperda and Aedes aegypti but did not improve the toxicity of Cry1C towards either insect pest. These results appeared to conflict with the aforementioned bioassay comparison between the EG6346 Cry1C protein and the aizawai 7.29 Cry1C protein showing no effect of amino acid substitutions within loop .beta. 6-7 of Cry1C on insecticidal specificity. Accordingly, the cry1C gene from strain aizawai 7.29 was re-sequenced where variant codons for the active toxin region were reported by Sanchis et al., (1989) and in PCT WO 88/09812. The results of that sequence analysis revealed no differences in the amino acid sequences of the active toxins of Cry1C from strain EG6346 and of Cry1C from strain aizawai 7.29. Thus, the prior art on the Cry1C protein of strain aizawai 7.29, in light of the aforementioned bioassay comparisons with the Cry1C protein of strain EG6346, incorrectly taught that multiple amino acid substitutions within loop .beta. 6-7 of Cry1C had no effect on insecticidal specificity. Recently, Smith et al., (1996) also reported unspecified sequencing errors in the aizawai 7.29 cry1C gene.
1.2.5 MOLECULAR GENETIC TECHNIQUES FACILITATE PROTEIN ENGINEERING
The revolution in molecular genetics over the past decade has facilitated a logical and orderly approach to engineering proteins with improved properties. Site specific and random mutagenesis methods, the advent of polymerase chain reaction (PCR.TM.) methodologies, and related advances in the field have permitted an extensive collection of tools for changing both amino acid sequence, and underlying genetic sequences for a variety of proteins of commercial, medical, and agricultural interest.
Following the rapid increase in the number and types of crystal proteins which have been identified in the past decade, researchers began to theorize about using such techniques to improve the insecticidal activity of various crystal proteins. In theory, improvements to .delta.-endotoxins should be possible using the methods available to protein engineers working in the art, and it was logical to assume that it would be possible to isolate improved variants of the wild-type crystal proteins isolated to date. By strengthening one or more of the aforementioned steps in the mode of action of the toxin, improved molecules should provide enhanced activity, and therefore, represent a breakthrough in the field. If specific amino acid residues on the protein are identified to be responsible for a specific step in the mode of action, then these residues can be targeted for mutagenesis to improve performance.
1.2.6 STRUCTURAL ANALYSES OF CRYSTAL PROTEINS
The combination of structural analyses of B. thuringiensis toxins followed by an investigation of the function of such structures, motifs, and the like has taught that specific regions of crystal protein endotoxins are, in a general way, responsible for particular functions.
For example, the structure of Cry3A (Li et al., 1991) and Cry1Aa (Grochulski et al., 1995) illustrated that the Cry1 and Cry3 .delta.-endotoxins have three distinct domains. Each of these domains has, to some degree, been experimentally determined to assist in a particular function. Domain 1, for example, from Cry3B2 and Cry1Ac has been found to be responsible for ion channel activity, the initial step in formation of a pore (Walters et al., 1993; Von Tersch et al., 1994). Domains 2 and 3 have been found to be responsible for receptor binding and insecticidal specificity (Aronson et al., 1995; Caramori et al., 1991; Chen et al. 1993; de Maagd et al., 1996; Ge et al., 1991; Lee et al., 1992; Lee et al., 1995; Lu et al., 1994; Smedley and Ellar, 1996; Smith and Ellar, 1994; Rajamohan et al., 1995; Rajamohan et al., 1996; Wu and Dean, 1996). Regions in domain 3 can also impact the ion channel activity of some toxins (Chen et al., 1993, Wolfersberger et al., 1996).
1.3 Defficiencies in the Prior Art
Unfortunately, while many laboratories have attempted to make mutated crystal proteins, few have succeeded in making mutated crystal proteins with improved lepidopteran toxicity. In almost all of the examples of genetically-engineered B. thuringiensis toxins in the literature, the biological activity of the mutated crystal protein is no better than that of the wild-type protein, and in many cases, the activity is decreased or destroyed altogether (Almond and Dean, 1993; Aronson et al., 1995; Chen et al., 1993, Chen et al., 1995; Ge et al., 1991; Kwak et al., 1995; Lu et al., 1994; Rajamohan et al., 1995; Rajamohan et al., 1996; Smedley and Ellar, 1996; Smith and Ellar, 1994; Wolfersberger et al., 1996; Wu and Aronson, 1992). For a crystal protein having approximately 650 amino acids in the sequence of its active toxin, and the possibility of 20 different amino acids at each of these sites, the likelihood of arbitrarily creating a successful new structure is remote, even if a general function to a stretch of 250-300 amino acids can be assigned. Indeed, the above prior art with respect to crystal protein gene mutagenesis has been concerned primarily with studying the structure and function of the crystal proteins, using mutagenesis to perturb some step in the mode of action, rather than with engineering improved toxins.
Several examples, however, do exist in the prior art where improvements to biological activity were achieved by preparing a recombinant crystal protein. Angsuthanasamnbat et al. (1993) demonstrated that a stretch of amino acids in the dipteran-toxic Cry4B delta-endotoxin is proteolytically sensitive and, by repairing this site, the dipteran toxicity of this protein was increased three-fold. In contrast, the elimination of a trypsin cleavage site on the lepidopteran-toxic Cry9C protein was reported to have no effect on insecticidal activity (Lambert et al., 1996). In another example, Wu and Dean (1996) demonsed that specific changes to amino acids at residues 481-486 (domain 2) in the coleopteran-toxic Cry3A protein increased the biological activity of this protein by 2.4-fold against one target insect, presumably by altering toxin binding. Finally, chimeric Cry1 proteins containing exchanges of domain 2 or domain 3 sequences and exhibiting improved toxicity have been reported, but there is no evidence that toxicity has been improved for more tand one lepidopteran insect pest or that insecticidal activity towards other lepidopteran pests has been retained (Caramori et al., 1991; Ge et al., 1991, de Maagd et al., 1996). Based on the prior art, exchanges involving domain 2 or domain 3 would be expected to change insecticidal specificity.
The prior art also provides examples of Cry1A mutants containing mutations encoding amino acid substitutions within the predicted .alpha. helices of domain 1 (Wu and Aronson, 1992; Aronson et al., 1995, Chen et al., 1995). None of these mutations resulted in improved insecticidal activity and many resulted in a reduction in activity, particularly those encoding substitutions within the predicted helix 5 (Wu and Aronson, 1992). Extensive mutagenesis of loop regions within domain 2 have been shown to alter the insecticidal specificity of Cry1C but to not improve its toxicity towards any one insect pest (Smith and Ellar, 1994). Similarly, extensive mutagenesis of loop regions in domain 2 and of .beta.-strand structures in domain 3 of the Cry1A proteins have failed to produce Cry1A mutants with improved toxicity (Aronson et al., 1995; Chen et al., 1993; Kwak et al., 1995; Smedley and Ellar, 1996; Rajamohan et al., 1995; Rajamohan et al., 1996). These results demonstrate the difficulty in engineering improved insecticidal proteins and illustrate that successful engineering of B. thuringiensis toxins does not follow simple and predictable rules.
Collectively, the limited successes in the art to develop synthetic toxins with improved insecticidal activity have stifled progress in this area and confounded the search for improved endotoxins or crystal proteins. Rather than following simple and predictable rules, the successful engineering of an improved crystal protein may involve different strategies, depending on the crystal protein being improved and the insect pests being targeted. Thus, the process is highly empirical.
Accordingly, traditional recombinant DNA technology is clearly not routine experimentation for providing improved insecticidal crystal proteins. What are lacking in the prior art are rational methods for producing genetically-engineered B. thuringiensis Cry1 crystal proteins that have improved insecticidal activity and, in particular, improved toxicity towards a wide range of lepidopteran insect pests.